Hello Mike Percy,
I'd like you to do a code review. Please visit
http://gerrit.cloudera.org:8080/4139
to review the following change.
Change subject: subprocess: use SIGKILL instead of SIGTERM death signal
......................................................................
subprocess: use SIGKILL instead of SIGTERM death signal
This addresses a potential source of test timeouts if a test starts a
minicluster, pauses one of the daemons with SIGSTOP, and then exits
unexpectedly (eg due to a CHECK failure). With the SIGTERM death signal,
I found that this resulted in the test hanging on exit: apparently it
delivered the SIGTERM to the stopped child and was then waiting on it
indefinitely.
After changing to SIGKILL I see the broken test failing from the FATAL
as expected (rather than timing out).
Change-Id: Icc411b8c031c1c117bdf0d82f272da2263c97d88
---
M src/kudu/util/subprocess.cc
1 file changed, 1 insertion(+), 1 deletion(-)
git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/39/4139/1
--
To view, visit http://gerrit.cloudera.org:8080/4139
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings
Gerrit-MessageType: newchange
Gerrit-Change-Id: Icc411b8c031c1c117bdf0d82f272da2263c97d88
Gerrit-PatchSet: 1
Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-Owner: Todd Lipcon <t...@apache.org>
Gerrit-Reviewer: Mike Percy <mpe...@apache.org>
